Document extractor service (#15665)

* Document extractor service

* Fixing vendor modules

* Addressing PR Review comments

* Some small simplifications

* Fixing a linter complain

* simplifying a bit the code using package variables

Co-authored-by: Mattermod <mattermod@users.noreply.github.com>

services/docextractor/documents.go

@@ -0,0 +1,61 @@
+// Copyright (c) 2015-present Mattermost, Inc. All Rights Reserved.
+// See LICENSE.txt for license information.
+
+package docextractor
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"path"
+	"strings"
+
+	"code.sajari.com/docconv"
+)
+
+type documentExtractor struct{}
+
+var doconvConverterByExtensions = map[string]func(io.Reader) (string, map[string]string, error){
+	"doc":   docconv.ConvertDoc,
+	"docx":  docconv.ConvertDocx,
+	"pptx":  docconv.ConvertPptx,
+	"odt":   docconv.ConvertODT,
+	"html":  func(r io.Reader) (string, map[string]string, error) { return docconv.ConvertHTML(r, true) },
+	"pages": docconv.ConvertPages,
+	"rtf":   docconv.ConvertRTF,
+}
+
+func (de *documentExtractor) Match(filename string) bool {
+	extension := strings.TrimPrefix(path.Ext(filename), ".")
+	_, ok := doconvConverterByExtensions[extension]
+	return ok
+}
+
+func (de *documentExtractor) Extract(filename string, r io.Reader) (string, error) {
+	extension := strings.TrimPrefix(path.Ext(filename), ".")
+	converter, ok := doconvConverterByExtensions[extension]
+	if !ok {
+		return "", errors.New("Unknown converter")
+	}
+
+	f, err := ioutil.TempFile(os.TempDir(), "docconv")
+	if err != nil {
+		return "", fmt.Errorf("error creating temporary file: %v", err)
+	}
+	defer f.Close()
+	defer os.Remove(f.Name())
+
+	_, err = io.Copy(f, r)
+	if err != nil {
+		return "", fmt.Errorf("error copying data into temporary file: %v", err)
+	}
+
+	text, _, err := converter(f)
+	if err != nil {
+		return "", err
+	}
+
+	return text, nil
+}